Application of detection reagent of five tryptophan metabolites in preparation of ER positive breast cancer early diagnosis kit

Abstract

The invention provides an application of a detection reagent of five tryptophan metabolites in preparation of an early diagnosis kit for ER (estrogen receptor) positive breast cancer. The tryptophan metabolites are kynurenine Kyn, indole-3-carboxylic acid glucuronide I3GA, indole-3-acetic acid I3A, indole-3-acetyl-L-alanine IAA-L-Ala and L-tryptophan L-Trp. According to the invention, early accurate diagnosis of ER positive breast cancer is realized, and the technical problems of insufficient efficiency of traditional diagnosis means and lack of'combined application 'in existing metabolite research are solved.

Description

Technical Field

[0001] This invention belongs to the field of biological detection and medical diagnostic technology, and relates to the application of detection reagents for five tryptophan metabolites in the preparation of early diagnostic kits for ER-positive breast cancer. Background Technology

[0002] ER-positive breast cancer (estrogen receptor positive) accounts for 60%-70% of all breast cancer cases. Its growth is estrogen-dependent, and early diagnosis is crucial for benefiting from endocrine therapy. However, current imaging examinations lack sensitivity for early lesions, and traditional plasma biomarkers (such as CA15-3 and CEA) have low positivity rates in early-stage patients and are easily interfered with by benign diseases. Therefore, a highly specific early diagnostic method for ER-positive breast cancer is urgently needed.

[0003] With the development of metabolomics technology, plasma metabolites have attracted widespread attention as potential biomarkers for disease diagnosis. Tryptophan, an essential amino acid, has a metabolic pathway closely related to the occurrence and development of tumors. Previous studies have shown that the levels of tryptophan and its metabolites in the plasma of breast cancer patients differ from those in healthy individuals; however, existing research has three major limitations:

[0004] ① The inclusion and exclusion criteria are vague, and underlying diseases such as diabetes and hypertension that affect tryptophan metabolism are not excluded, resulting in non-tumor factors mixed with differential metabolites;

[0005] ② The sample characteristics are limited, mostly including newly diagnosed and untreated patients, and do not cover different stages, grades, treatment statuses and relapse populations, which cannot meet the needs of the entire diagnosis and treatment process;

[0006] ③ Existing studies mostly focus on single metabolites and have failed to construct a "combination of metabolic markers" that is highly specific and discriminative for ER-positive breast cancer, thus limiting their clinical value. Summary of the Invention

[0007] In view of this, the purpose of the present invention is to provide the application of detection reagents for five tryptophan metabolites in the preparation of early diagnostic kits for ER-positive breast cancer.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] 1. Application of detection reagents for five tryptophan metabolites in the preparation of an early diagnostic kit for ER-positive breast cancer, wherein the tryptophan metabolites are kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), indole-3-acetic acid (I3A), indole-3-acetyl-L-alanine (IAA-L-Ala), and L-tryptophan (L-Trp).

[0010] Preferably, the test sample for the kit is plasma.

[0011] 2. An ER-positive breast cancer early diagnostic kit, comprising detection reagents for five tryptophan metabolites, wherein the tryptophan metabolites are kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), indole-3-acetic acid (I3A), indole-3-acetyl-L-alanine (IAA-L-Ala), and L-tryptophan (L-Trp).

[0012] 3. A method for detecting plasma samples for non-disease diagnostic and therapeutic purposes, the specific steps of which are as follows:

[0013] (1) Pretreatment of plasma samples: Add mixed extraction solvent and isotope internal standard solution to plasma samples, vortex, sonicate, incubate at low temperature, centrifuge to obtain supernatant I, freeze dry under vacuum, reconstitute with reconstitution solvent, centrifuge to obtain supernatant II, filter, and inject for detection;

[0014] (2) LC-MS / MS detection;

[0015] (3) Plot the standard curve and calculate the concentrations of the five tryptophan metabolites in the plasma sample;

[0016] The tryptophan metabolites mentioned are kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), indole-3-acetic acid (I3A), indole-3-acetyl-L-alanine (IAA-L-Ala), and L-tryptophan (L-Trp).

[0017] Preferably, in step (1), the volume ratio of plasma sample, mixed extraction solvent, and isotope internal standard solution is 50:200:2.5. The mixed extraction solvent is obtained by mixing acetonitrile, methanol, and formic acid in a volume ratio of 50:50:0.1. The isotope internal standard solution contains tryptophan-d5 (Trp-d5, 100 nM), melatonin-d4 (MEL-d4, 20 nM), 5-fluorotryptamine (5-FT, 25 nM), indole-3-acetic acid-d5 (IAA-d5, 50 nM), 4-chloro-phenylalanine (4-Cl-Phe, 50 nM), and 1-undecaneamine (1-UA, 50 nM).

[0018] Preferably, in step (1), the vortex oscillation conditions are: 3000rpm vortex oscillation for 5min; the ultrasonic treatment conditions are: 300W, 40kHz ultrasonic treatment for 10min; and the low temperature incubation conditions are: -80℃ low temperature incubation for 30min.

[0019] Preferably, in step (1), the volume ratio of supernatant I to the resolution solvent is 9:5, and the resolution solvent is obtained by mixing methanol, water and formic acid in a volume ratio of 50:50:0.1.

[0020] Preferably, in step (2), chromatographic separation is performed on an Agilent 1290 II system using a ZORBAX EclipsePlus C18 column; mass spectrometry analysis is performed on a SCIEX Triple Quad™ 6500 plus QTRAP in negative ion electrospray mode.

[0021] Preferably, the specific method of step (3) is as follows: prepare a standard concentration gradient (from 0.05 nM to 100000 nM), add internal standard solution to each concentration, detect by LC-MS / MS, plot the standard curve with the standard concentration as the abscissa and the standard peak area / internal standard peak area as the ordinate, fit the regression equation using the weighted least squares method, and calculate the concentration of each metabolite in the plasma sample according to the regression equation.

[0022] The beneficial effects of this invention are as follows:

[0023] This invention provides the application of detection reagents for five tryptophan metabolites in the preparation of an early diagnostic kit for ER-positive breast cancer. The tryptophan metabolites are kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), indole-3-acetic acid (I3A), indole-3-acetyl-L-alanine (IAA-L-Ala), and L-tryptophan (L-Trp).

[0024] This invention eliminates interference from underlying diseases through strict inclusion and exclusion criteria, and combines standardized LC-MS / MS detection technology to systematically analyze the differences in plasma metabolism between healthy individuals and ER-positive breast cancer patients. It screens and determines novel combinations of tryptophan metabolites for breast cancer diagnosis, enabling early and accurate diagnosis of ER-positive breast cancer. This solves the technical problems of insufficient efficacy of traditional diagnostic methods and the lack of "combined application" in existing metabolite research.

[0025] Specifically, this study utilized liquid chromatography-tandem mass spectrometry (LC-MS / MS) to detect tryptophan and its 24 metabolites in plasma, and screened and determined a novel combination of tryptophan metabolites as diagnostic markers for ER-positive breast cancer. Through strict inclusion and exclusion criteria, 88 healthy individuals and 109 ER-positive breast cancer patients (covering different stages, grades, and pre-treatment states) were screened. Differences in plasma metabolism were systematically analyzed, ultimately identifying a combination of tryptophan metabolites with high diagnostic efficacy, applicable to early screening, efficacy evaluation, and recurrence monitoring of ER-positive breast cancer.

[0026] This invention has the following advantages:

[0027] 1. Diagnostic markers have high specificity and clear biological significance.

[0028] This invention identifies IAA-L-Ala and I3GA as core differential metabolites in ER-positive breast cancer through a rigorous dual screening criterion (P<0.05 and |fold change|>2). Both are key products of the tryptophan metabolism pathway, and changes in their expression levels directly reflect the metabolic disorder characteristics of ER-positive breast cancer patients.

[0029] 2. The combination of multiple biomarkers offers excellent diagnostic efficacy, balancing accuracy and practicality.

[0030] An innovative combination of tryptophan metabolite diagnostic biomarkers was proposed: breaking through the limitations of existing research on "single metabolites", five tryptophan metabolites, namely Kyn, I3GA, I3A, IAA-L-Ala, were screened for the first time. The diagnostic AUC (0.913) can be used as a biomarker for early diagnosis of ER-positive breast cancer.

[0031] In the multi-marker combination diagnostic model constructed in this invention, the AUC of the combination of 5 diagnostic markers reached 0.913 (95% confidence interval 0.859-0.965), with a prediction accuracy of 83.2%. This is significantly better than the diagnostic efficacy of single metabolites (I3GA AUC 0.803, IAA-L-Ala AUC 0.647), while avoiding the problems of high detection cost and complex operation of high-number marker combinations (such as 23 markers). It achieves a balance between diagnostic performance and clinical practicality and is suitable for large-scale screening and auxiliary diagnosis of ER-positive breast cancer.

[0032] 3. Standardized sample processing and testing methods ensure reliable and reproducible results.

[0033] This invention establishes a standardized procedure for plasma sample collection, preprocessing, and LC-MS / MS detection: by strictly defining sample inclusion and exclusion standards to eliminate interfering factors, using isotope internal standard method and weighted least squares method to fit standard curves, and combining left-truncated data estimation method to handle missing values, median normalization and automatic scaling combination normalization strategy, effectively eliminating experimental systematic errors and data bias, ensuring the stability and repeatability of test results, and providing technical support for clinical promotion.

[0034] Strict inclusion and exclusion criteria ensure biomarker specificity: The system systematically excludes more than 10 types of underlying diseases that interfere with tryptophan metabolism, such as metabolic, cardiovascular, hepatobiliary, and thyroid diseases. At the same time, it only includes patients before treatment to avoid the influence of treatment on metabolites and ensure that the differences in the combination of 5 metabolites are only associated with ER-positive breast cancer, with a specificity of 90.9%.

[0035] Standardized testing techniques ensure reliable results: The Agilent 1290 II + SCIEX 6500 plus QTRAP combined system is used, and the negative electrospray ionization mode is adapted to the acidic characteristics of tryptophan metabolites; the methodology validation fully complies with FDA / EMA guidelines, and the matrix effect, precision, stability and other indicators all meet the standards, ensuring that the test results are reproducible and can be widely applied.

[0036] 4. The diagnostic model has a complete validation system and high clinical translational value.

[0037] This invention constructs a complete diagnostic efficacy verification system of "univariate validation-multivariate optimization". Through multiple verification methods such as ROC curve analysis, cross-validation prediction probability distribution, and confusion matrix, it comprehensively confirms the diagnostic reliability of biomarker combinations. At the same time, it clarifies the optimal critical values ​​and clinical fit scenarios for each biomarker combination, providing a direct basis for the development of subsequent diagnostic kits and the setting of clinical testing thresholds, and has significant industrialization prospects.

[0038] Blinding and multi-dimensional validation enhance credibility: Clinical information blinding is used throughout the sample testing process to avoid subjective bias; the combination of biomarkers is determined through a three-step method of "differential screening → Logistic modeling → ROC validation", and the diagnostic value of the combination is validated from methodological and statistical perspectives. Attached Figure Description

[0039] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration:

[0040] Figure 1 This is a flowchart of the present invention.

[0041] Figure 2 This method uses left-truncated data estimation to handle missing values.

[0042] Figure 3 To normalize the data.

[0043] Figure 4 To screen for differentially expressed metabolites, a threshold of 2 was set for fold change.

[0044] Figure 5 To screen for differentially expressed metabolites, a t-test p-value threshold of 0.05 was set.

[0045] Figure 6 To screen differentially expressed metabolites, the dual criteria of P<0.05 and |FC|>2 were met.

[0046] Figure 7The difference in plasma concentrations of I3GA and IAA-L-Ala in ER-positive breast cancer patients and healthy controls is represented by A, where A represents the plasma concentration of I3GA in the two groups and B represents the plasma concentration of IAA-L-Ala in the two groups.

[0047] Figure 8 The diagnostic value of I3GA and IAA-L-Ala was analyzed using classic univariate ROC curves, where A is the ROC curve of I3GA and B is the ROC curve of IAA-L-Ala.

[0048] Figure 9 To explore the diagnostic value of different combinations of metabolites based on multivariate ROC curves, where A represents the multivariate ROC curves of different diagnostic biomarker combination models, and B represents the prediction accuracy of different diagnostic biomarker combination models.

[0049] Figure 10 The optimal model for the diagnostic models of five tryptophan metabolites was determined, where A is the compositional feature of the five tryptophan metabolite diagnostic models and B is the sample predicted class probability using the optimal classifier. Detailed Implementation

[0050] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0051] The process of this invention is as follows Figure 1 As shown.

[0052] Example 1

[0053] Sample collection and preprocessing

[0054] Health check-up group (88 people): Class 1

[0055] Inclusion criteria: ① Female, aged 30-72 years; ② Normal levels of complete blood count, liver and kidney function, fasting blood glucose, blood lipids, thyroid function, tumor markers (CA15-3, CEA), and other indicators; ③ No history of breast disease or malignant tumors.

[0056] Exclusion criteria: Individuals who meet any of the following "Exclusion criteria for breast cancer group".

[0057] Breast cancer group (109 cases): Class 0

[0058] Inclusion criteria: ① Age 30-73 years, all female, pathologically confirmed as ER-positive breast cancer; ② Basic complete clinical data, including tumor TNM stage (19 cases of stage I, 46 cases of stage II, 7 cases of stage III, 7 cases of stage IV, and 30 cases with no clear stage) and molecular subtype (4 cases of Luminal A, 47 cases of Luminal B, 18 cases of HER2, and 40 cases with no clear subtype).

[0059] Exclusion criteria: ① Metabolic diseases: complicated with hypoglycemia, impaired glucose tolerance, diabetes, hypercholesterolemia, or hyperuricemia; ② Cardiovascular and cerebrovascular diseases: complicated with hypertension, arteriosclerosis, coronary heart disease, cerebral infarction, or other organic heart diseases; ③ Hepatobiliary diseases: complicated with viral hepatitis, liver cysts, fatty liver, liver dysfunction, drug-induced liver injury, gallstones, or adenomyoma of the gallbladder; ④ Thyroid diseases: complicated with hyperthyroidism, hypothyroidism, nodular goiter, thyroid tumors, parathyroid adenomas, or thyroiditis; ⑤ Other tumors: complicated with primary or metastatic tumors in other organs such as the liver, lungs, kidneys, adrenal glands, or bladder; ⑥ Electrolyte disorders: complicated with hyperkalemia or hypokalemia; ⑦ Mental illnesses: complicated with depression.

[0060] Sample inclusion and exclusion: 88 healthy women undergoing physical examinations (aged 30-72 years, with normal physical examination indicators and no history of breast disease or malignant tumors) were selected as the healthy control group (Group 1); 109 female patients with pathologically confirmed breast cancer (aged 30-73 years, with complete clinical data) were selected as the ER-positive breast cancer group (Group 0), excluding patients with confounding factors such as metabolic diseases, cardiovascular and cerebrovascular diseases, and hepatobiliary diseases.

[0061] Sample collection and preservation: 5 mL of venous blood was collected from the subjects on an empty stomach into heparin anticoagulant tubes. The plasma was separated by centrifugation at 3000 rpm for 10 min, aliquoted into 200 μL / tube, and stored in an ultra-low temperature freezer at -80℃. Repeated freeze-thaw cycles should be avoided (≤2 times) to ensure the stability of metabolites.

[0062] Example 2

[0063] LC-MS / MS detection and standard curve plotting

[0064] Detection conditions: Compounds were separated using an Agilent 1290 LC system. The column was an Agilent Technologies InfinityLab Poroshell 120 EC-C18 column (2.1 × 100 mm, 2.7 μm particle size), manufactured in California, USA. The column temperature was maintained at 40 °C. The mobile phase consisted of mobile phase A (water / formic acid, v / v) and mobile phase B (acetonitrile / formic acid, v / v). The elution program was set as follows: initially, mobile phase A was 95% concentrated and held for 0.25 min; then linearly decreased to 5% within 5.75 min and held for 2 min; then increased back to 95% mobile phase A within 0.1 min, and finally equilibrated for 1.9 min before the next injection. The flow rate was 0.4 mL / min, and the injection volume was 5 μL. The samples were placed in an autosampler and stored at 4 °C throughout the process.

[0065] Mass spectrometry analysis was performed using a SCIEX Triple Quad™ 6500 plus QTRAP mass spectrometer (SCIEX Triple Quad™ 6500 plus QTRAP) from Thermo Fisher Scientific (Massachusetts, USA) to detect analyte signal intensity. The ion source was an electrospray ionization source, and the scanning mode was multiple reaction monitoring (MRM) with positive / negative ion polarity switching. The MRM detection window for each compound was 60 s, and the total duration of a single positive / negative ion polarity switching cycle was 0.65 s. The specific precursor ion (Q1), daughter ion (Q3), declustering voltage (DP), collision energy (CE), and collision chamber exit voltage (CXP) in MRM were optimized by directly injecting a 500 nM standard solution at a flow rate of 7 μL / min to screen for optimal conditions. In positive ion mode, the ion spray voltage was 5500 V and the ion source temperature was 550 °C; in negative ion mode, the ion spray voltage was 4500 V and the ion source temperature was 450 °C. The flow rates of the curtain gas, ion source gas 1, and ion source gas 2 were 40 L / h, 50 L / h, and 50 L / h, respectively, and the collision gas intensity was set to medium.

[0066] Standard curve preparation: Prepare a concentration gradient of tryptophan metabolite standards (0.05, 0.5, 5, 10, 50, 100, 500, 1000, 5000, 10000, 50000, 10000, 100000 nM). Add 2.5 μL of internal standard solution containing tryptophan-d5 (Trp-d5, 100 nM), melatonin-d4 (MEL-d4, 20 nM), 5-fluorotryptamine (5-FT, 25 nM), indole-3-acetic acid-d5 (IAA-d5, 50 nM), 4-chlorophenylalanine (4-Cl-Phe, 50 nM), and 1-undecaneamine (1-UA, 50 nM) according to the pretreatment and detection procedures described above. Repeat each concentration three times and take the average peak area. Using "standard concentration (x)" as the x-axis and "standard peak area / internal standard peak area (y)" as the y-axis, a weighted least squares method (weight 1 / x²) was used to fit the regression equation, requiring R² > 0.995 and residuals < 15% at each concentration point. The concentrations of each metabolite in the sample were calculated based on the regression equation.

[0067] Plasma sample analysis: 50 μL of plasma sample was taken and 200 μL of acetonitrile / methanol / formic acid mixed solution (volume ratio 50:50:0.1) was added for metabolite extraction. Simultaneously, 2.5 μL of internal standard solution containing Trp-d5 (2 μM), MEL-d4 (400 nM), 5-FT (500 nM), IAA-d5 (1 μM), 4-Cl-Phe (1 μM), and 1-UA (1 μM) was added. After vigorous vortexing for 20 s, the mixture was sonicated in ice water for 10 min, followed by incubation at 80 ℃ for 30 min. The mixture was then centrifuged at 4 ℃ and 14000 × g for 10 min. The supernatant was transferred to a clean centrifuge tube and dried using a vacuum centrifuge concentrator. Add 50 μL of a methanol / water / formic acid mixture (50:50:0.1 v / v) to the dried residue to reconstitute it, then sonicate it again in ice water for 10 min, and centrifuge it at 4 ℃ and 14000 ×g for 10 min. The resulting supernatant is ready for instrumental analysis. All sample tests were performed blinded, and the testers were not informed of the corresponding clinical information of the samples to avoid subjective bias.

[0068] Example 3

[0069] Missing value handling and data normalization

[0070] During data processing, 13 data points were missing in the tumor group. These missing data points would hinder subsequent downstream analysis and affect the accuracy and reliability of the results. To address this, the applicant adopted a left-truncated data estimation method, specifically as follows: 1 / 5 of the smallest positive value of each feature was used as a substitute value for its detection limit, and this substitute value was used to replace the aforementioned missing data points. Figure 2 ).

[0071] To eliminate data system bias and optimize downstream analysis results, the metabolite quantification data obtained by LC-MS / MS detection were further normalized. The applicant combined the following two methods to achieve better data correction results, specifically including:

[0072] Sample normalization: used to make general adjustments for systematic differences between samples. This invention adopts median normalization, which uses the median of all features in the sample as a benchmark for data calibration. This effectively offsets systematic errors caused by factors such as instrument status and experimental conditions during sample detection, and ensures that the data between different samples are comparable.

[0073] Data scaling: The scaling factor is calculated based on the dispersion of variables / features, and the scale is adjusted for each feature. This invention uses the auto scaling method, which first performs mean centering on each feature data (subtracting the mean of the feature), and then divides it by the standard deviation of the feature, so that the processed data meets the standard distribution characteristics of mean 0 and variance 1. This not only preserves the distribution information of the feature itself, but also eliminates the weight imbalance problem caused by the difference in units or numerical ranges between different features.

[0074] By combining the two methods described above, the dual problems of systematic differences between samples and scale inconsistencies between features can be solved simultaneously, laying a reliable data foundation for subsequent screening of differential metabolites. Figure 3 ).

[0075] Example 4

[0076] Differential metabolite screening

[0077] Preliminary screening of differentially expressed metabolites: A dual screening criterion was used to analyze the differences in metabolites between the plasma of ER-positive breast cancer patients and the control group. A fold change threshold of 2 was set, identifying two significantly upregulated metabolites (indole-3-acetyl-L-alanine, IAA-L-Ala; 3-indoleacetic acid-O-glucuronic acid, I3GA) and one significantly downregulated metabolite (indole) in the plasma of breast cancer patients. Figure 4The t-test p-value threshold was set to 0.05, and 13 differentially metabolites were identified: kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), 5-hydroxyindoleacetic acid (5HIAA), L-tryptophan (L-Trp), kynurenine acid (KynA), indole-3-acetic acid (I3A), quinolinic acid (QA), indole-3-acetyl-L-alanine (IAA-L-Ala), indole-3-carboxylic acid, phthalimide acetic acid (PIC), xanthurenic acid, indolelactic acid (ILA), and indoleacetic acid (IAA). Figure 5 ).

[0078] Dual-criteria screening: This invention screens for core differentially expressed metabolites—IAA-L-Ala and I3GA—that are significantly and specifically upregulated in the plasma of breast cancer patients by simultaneously satisfying the dual criteria of P<0.05 and |fold change (FC)|>2. Figure 6 The screening results demonstrated both statistical significance and biological significance, effectively eliminating the interference of false-positive differential metabolites. They clarified the association between the two metabolites and the development of breast cancer, providing highly specific and sensitive potential metabolic biomarkers for the early diagnosis and disease monitoring of ER-positive breast cancer. This constitutes the core technical advantage of this invention in the differential metabolite screening process. Figure 7 ).

[0079] Example 5

[0080] Diagnostic model construction and validation

[0081] To comprehensively validate the diagnostic value of differentially expressed metabolites and construct a better diagnostic model, this invention sequentially conducts classical univariate ROC curve analysis and exploratory analysis based on multivariate ROC curves, forming a complete diagnostic efficacy validation system of "univariate validation - multivariate optimization". The specific technical solution is as follows:

[0082] Classical univariate ROC curve analysis: The core objective is to verify the diagnostic differentiation ability of individual differentially expressed metabolites for breast cancer. Specific procedures include: constructing univariate ROC curves for each screened differentially expressed metabolite, calculating the area under the curve (AUC) or partial AUC and its 95% confidence interval, and quantifying the diagnostic efficacy of each metabolite in differentiating breast cancer patients from healthy controls. This invention clarifies the diagnostic differentiation ability of the core differentially expressed metabolites IAA-L-Ala and I3GA through this analysis. The AUC of I3GA is 0.803, with a 95% confidence interval of 0.734-0.864; the AUC of IAA-L-Ala is 0.647, with a 95% confidence interval of 0.572-0.727. Neither of their 95% confidence intervals include 0.5, confirming their independent and significant diagnostic differentiation ability. Performance data at different thresholds provide quantitative support for selecting differentiated diagnostic strategies in clinical scenarios (such as screening scenarios emphasizing high sensitivity or diagnostic scenarios emphasizing high specificity). Figure 8 ).

[0083] Exploratory Analysis Based on Multivariate ROC Curves: To overcome the limitations of univariate analysis, which can only extract the value of a single metabolite, and to further improve diagnostic accuracy and robustness, this invention conducts this analysis to achieve automated identification of important features of differentially expressed metabolites and joint diagnostic performance evaluation. The specific technical solution is as follows: Linear Support Vector Machine (SVM) is used as the classification method. This method has excellent adaptability and generalization ability to high-dimensional data and can accurately capture the linear correlation between metabolite features and breast cancer disease status. The feature ranking method adopts the SVM built-in ranking mechanism, which can quantify the contribution of each differentially expressed metabolite to the classification results, realizing automated screening and priority ranking of key diagnostic features. This invention obtained specific diagnostic efficacy data for combined diagnostic models with different combinations of diagnostic biomarkers through this analysis. Specifically: the AUC of the model with 2 diagnostic biomarkers was 0.874, with a 95% confidence interval of 0.757-0.936; the AUC of the model with 3 diagnostic biomarkers was 0.898, with a 95% confidence interval of 0.855-0.940; the AUC of the model with 5 diagnostic biomarkers was 0.913, with a 95% confidence interval of 0.859-0.965; the AUC of the model with 10 diagnostic biomarkers was 0.929, with a 95% confidence interval of 0.886-0.969; the AUC of the model with 20 diagnostic biomarkers was 0.945, with a 95% confidence interval of 0.895-0.989; and the AUC of the model with 23 diagnostic biomarkers was 0.946, with a 95% confidence interval of 0.900-0.990. Figure 9(A). This analysis achieved two key results: First, the AUC values ​​of the combined diagnostic model were significantly higher than those of the single metabolite (I3GA AUC 0.803, IAA-L-Ala AUC 0.647). Furthermore, the AUC of the combined model gradually increased with the number of diagnostic biomarkers, reaching a maximum of 0.946, confirming the significant superiority of multi-metabolite combined diagnosis and greatly improving the discriminative power of ER-positive breast cancer diagnosis. Second, through SVM-built-in feature ranking, the core contribution of IAA-L-Ala and I3GA in the combined model was further verified, strengthening their specificity and importance as core diagnostic biomarkers. In addition, the automated feature recognition process significantly improved analytical efficiency and reduced human error compared to traditional manual screening, providing efficient and accurate technical support for the development of subsequent multi-biomarker combined diagnostic kits.

[0084] Analysis of the predictive accuracy of different diagnostic biomarker combinations showed that the predictive accuracy of all multi-biomarker combination diagnostic models was higher than 80%; among them, the model with 23 diagnostic biomarker combinations achieved the best predictive accuracy, reaching 86.3%. Considering core factors such as clinical application feasibility, testing cost, and diagnostic efficacy, this invention determined 5 diagnostic biomarker combinations as the optimal practical solution—this combination model has an AUC value greater than 0.9 (0.913) and a predictive accuracy of 83.2%, balancing diagnostic performance with the convenience and economy of clinical testing, making it more suitable for clinical translational applications. Figure 9 (B)

[0085] Optimal Model Determination: Considering both diagnostic efficacy and clinical applicability, five biomarker combinations were selected as the optimal diagnostic model. Based on the optimal classifier (selected using AUC as the core evaluation metric), the average predicted class probability distribution of the five diagnostic biomarker combination models after cross-validation was presented (classification boundary x=0.5). During the model training phase, a balanced subsampling method was used, with the classification boundary kept constant at the center threshold (x=0.5, marked with a dashed line). The model's true positive, false positive, true negative, and false negative classification results were visually presented, further validating the diagnostic reliability and stability of the five diagnostic biomarker combination models from the perspective of classification results. Figure 10 ).

[0086] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. The application of detection reagents for five tryptophan metabolites in the preparation of an early diagnostic kit for ER-positive breast cancer, characterized in that, The tryptophan metabolites mentioned are kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), indole-3-acetic acid (I3A), indole-3-acetyl-L-alanine (IAA-L-Ala), and L-tryptophan (L-Trp).

2. A diagnostic kit for early diagnosis of ER-positive breast cancer, characterized in that, A detection reagent containing five tryptophan metabolites, namely kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), indole-3-acetic acid (I3A), indole-3-acetyl-L-alanine (IAA-L-Ala), and L-tryptophan (L-Trp).

3. A method for detecting plasma samples for non-disease diagnostic and therapeutic purposes, characterized in that, The specific steps are as follows: (1) Pretreatment of plasma samples: Add mixed extraction solvent and isotope internal standard solution to plasma samples, vortex, sonicate, incubate at low temperature, centrifuge to obtain supernatant I, freeze dry under vacuum, reconstitute with reconstitution solvent, centrifuge to obtain supernatant II, filter, and inject for detection; (2) LC-MS / MS detection; (3) Plot the standard curve and calculate the concentrations of the five tryptophan metabolites in the plasma sample; The tryptophan metabolites mentioned are kynurenine (Kyn), indole-3-carboxylic acid glucuronide (I3GA), indole-3-acetic acid (I3A), indole-3-acetyl-L-alanine (IAA-L-Ala), and L-tryptophan (L-Trp).

4. The detection method according to claim 3, characterized in that, In step (1), the volume ratio of plasma sample, mixed extraction solvent, and isotope internal standard solution is 50:200:2.

5. The mixed extraction solvent is obtained by mixing acetonitrile, methanol, and formic acid in a volume ratio of 50:50:0.

1. The isotope internal standard solution contains tryptophan-d5, melatonin-d4, 5-fluorotryptamine, indole-3-acetic acid-d5, 4-chloro-phenylalanine, and 1-undecaneamine.

5. The detection method according to claim 3, characterized in that, In step (1), the vortex oscillation conditions are: vortex oscillation at 3000 rpm for 5 min; the ultrasonic treatment conditions are: ultrasonic treatment at 300W and 40kHz for 10 min; and the low-temperature incubation conditions are: low-temperature incubation at -80℃ for 30 min.

6. The detection method according to claim 3, characterized in that, In step (1), the volume ratio of supernatant I to the resolution solvent is 9:

5. The resolution solvent is obtained by mixing methanol, water and formic acid in a volume ratio of 50:50:0.

1.

7. The detection method according to claim 3, characterized in that, In step (2), chromatographic separation was performed on an Agilent 1290II system using a ZORBAX Eclipse Plus C18 column; mass spectrometry analysis was performed on a SCIEX Triple Quad™ 6500 plus QTRAP in negative ion electrospray mode.

8. The detection method according to claim 3, characterized in that, The specific method of step (3) is as follows: prepare a standard concentration gradient, add internal standard solution to each concentration, detect by LC-MS / MS, plot the standard curve with the standard concentration as the abscissa and the standard peak area / internal standard peak area as the ordinate, fit the regression equation using the weighted least squares method, and calculate the concentration of each metabolite in the plasma sample according to the regression equation.

Images